The incidence of melanoma is rising, with the majority of cases diagnosed at localized stages, with relatively high cure rates [1]. However, recurrent and metastatic melanoma is associated with a worse prognosis. The emergence of immune checkpoint inhibitors have ushered in a new era of therapy for recurrent and advanced melanoma and many other [2–4]. In early 2011, the Food and Drug Administration (FDA) approved ipilimumab, an antibody that blocks the inhibitory receptor CTLA-4 expressed on T cells, (the first immunotherapeutic drug in the class of immune checkpoint inhibitors) for the treatment of advanced stage melanoma [5]. Antibodies directed against another inhibitory receptor, programmed death 1 (PD-1) and PD-1 ligand either used as monotherapy or in combination with ipilimumab have demonstrated overall survival benefit compared to ipilimumab alone and chemotherapy and are now approved by regulatory agencies and standard of care for the treatment of a number of solid and hematologic malignancies including melanoma [3, 4].

While retrospective studies have confirmed the survival benefit with immune checkpoint inhibitors in the treatment of metastatic melanoma observed in prospective studies, [6] there are scarce data on the adoption of immunotherapy in the community. We therefore aimed to investigate the use of immunotherapy for metastatic melanoma across hospitals over time, and sought to identify factors associated with the receipt of immunotherapy in the community. We hypothesized that certain hospitals are better equipped than others to adopt these new therapies.

We herein demonstrate not only how the use of immunotherapy for metastatic melanoma has spread over time but also how its implementation has varied across hospitals and what factors predict treatment at hospitals with low vs. high use of immunotherapy. Since the approval of ipilimumab as the first immunotherapeutic drug of its kind in 2011, immunotherapy has rapidly evolved and now represents first or second-line therapy for a variety of cancers [14, 15]. However, as demonstrated by our finding of significant facility-level variation in immunotherapy uptake, it is conceivable that the enormous economic burden of this new therapy [16] is hampering comprehensive implementation across hospitals.

When considering the general use of immunotherapy from the time of its first approval in 2011 to recent years, we found a gradual uptake in the use of immunotherapy across hospitals (Fig. 3) that is consistent with adoption curves witnessed with other novel drugs or devices [17]. The proportion of hospitals treating at least 20% of their patients with immunotherapy for metastatic melanoma within 90 days of diagnosis was approximately 2.5 times higher in 2015 compared to 2011. This trend is likely to continue as familiarity with targeted therapies increases among healthcare professionals [18].

Despite level-one evidence demonstrating a survival benefit associated with the use of immunotherapy in the treatment of metastatic melanoma, we noted significant facility-level variation in immunotherapy uptake [5]. Facility-level rates of immunotherapy use in high-prescribing hospitals approached 50%, compared to just 8% among low prescribing hospitals. Our results corroborate results from investigations regarding variations in the use of new therapeutics in other cancers [19]. Collectively, these results suggest that non-clinical predictors of care such as facility type may be contributing to care inequity that disproportionately affects underserved communities. Non-adherence to clinical guidelines and recommendations is a phenomenon that has repeatedly been shown across a variety of specialties and conditions (including melanoma), [20, 21] which in turn may affect clinical prognosis [22, 23]. Consequently, it is critical that providers and policymakers alike identify and eliminate drivers of healthcare that is either not indicated or inadequate.

Patient and physician-level factors must also be considered as a source of the variation observed in our study [20]. A lack of experience and poor access to information regarding the appropriate use of immunotherapy may discourage physician uptake, particularly given that immune-related toxicities can result in mortality and their management often requires specific expertise [24]. From the patient’s perspective, compliance with these novel drugs, especially in the context of adverse effects, requires adequate financial stability, as well as family/social support. Similarly, low prescribing hospitals were more likely to be non-academic centers that may not have early access to immunotherapy in the context of clinical trials which precede FDA approval and wider access to new agents. More than 80% of hospitals treating the highest proportion of patients with immunotherapy were academic. These academic institutions have greater access to clinical trials that may provide immunotherapy before FDA approval. Access to drugs in a clinical trial setting is likely to facilitate rapid implementation and routine use of new drugs after FDA approval because physicians will have greater familiarity with managing immune-related toxicities.

Financial aspects potentially affecting the care setting for metastatic melanoma patients must also be considered as evident by our finding that underinsured patients with Medicaid insurance or no insurance had a much higher probability of being treated in a low prescribing hospital. While drug coverage (as provided by Medicaid) is one aspect of the question, there are other factors around the treatment of the patient including payments to providers and hospitals that will be impacted by patient insurance. While most providers and hospitals – at least deliberately – do not select patients according to their insurance for the simple goal of maximizing profit, there is certainly a larger scale systemic incentive to do so. Our findings are consistent with prior work showing that underserved populations experience lower quality care across a variety of health care settings [25, 26]. The cost-intensive nature of immunotherapy is likely to exacerbate already observed health inequities experienced by the socioeconomically disadvantaged as hospitals and patients with lower means to pay for adequate treatment and lack of resources may affect treatment uptake and adherence [27]. Indeed, the administration of novel immunotherapy requires supplemental resources; in addition to the costs for the drug itself, there are added expenditures related to implementing support and pharmacy teams are required to correctly treat patients that are more easily borne by large academic centers.

Interestingly, the only clinical factor associated with lower odds of being treated at a low-prescribing hospital was the presence of visceral metastatic disease. However, factors classically used to define patient eligibility to systematic treatment, such as age or comorbidities, [28] were not different between the hospitals. There is evidence that better outcomes can be achieved when patients with complex diseases receive care at more specialized hospitals, supporting the concept of centralization [29]. It is possible that care for patients with more advanced disease may be more likely to be transferred to more experienced hospitals, there is no other clinical factor explaining differences in the use of immunotherapy.

We acknowledge that our work has some limitations. First, we are unable to adjust for intrinsic confounding given the retrospective observational nature of our study. Second, the database we used, NCDB, is a hospital-based registry that contains only information on patients treated at CoC-accredited hospitals. Our results may therefore not be representative for patients being treated outside of these facilities. Third, the NCDB does not capture the type or dosage of immunotherapy administered and approvals of PD-1/PD-L1 inhibitors fall in the latter time frame of our investigation. As a result, our data are more likely to reflect adoption of ipilimumab than adoption of nivolumab and pembrolizumab though we cannot distinguish use of individual immunotherapy agents. For the same reason, it is possible that some patients received experimental immunotherapy agents on clinical trials that were not FDA approved at the time of their administration. Although it is beyond the scope of our current investigation, it will be crucial to expand our next analysis to the timeframe between 2015 and 2018 to explore the broadening indications for immunotherapy. Greater familiarity with these agents with time may lead to more rapid adoption of immunotherapy in the community and increased use in non-academic centers.

While the use of immunotherapy for metastatic melanoma has increased over time, adoption varies widely across hospitals. Underinsured patients were more likely to receive treatment at low immunotherapy prescribing hospitals. The variation suggests inequity in access to these potentially life-saving drugs.